Without question, data is a critical corporate asset. That's why, to varying degrees, most enterprises have already become data-driven businesses. These days, leading organizations in all industries are expanding the scope of data they include in business processes and decision-making. One emerging category is called alternative data (or new data). Those who are willing to both reconsider their existing data – and consider alternative data – are likely to see significant, positive changes in the business.

How risky is this business of alternative data? Let's take a look.

What is alternative data?

Traditional data sources that businesses have used for decades often represent an enterprise’s digital intellectual property. Or, they may be the byproduct of internal business processes along with the data management and data governance functions that were built around them. In this sense, traditional data "feels" relatively safe.

Alternative data, on the other hand, is from nontraditional, mostly external sources. Consider, for example, data feeds from government agencies, industry associations and third-party data providers. Enterprises often collect this data directly from public websites through web scraping or web crawling. Some of those websites even provide public APIs to make their data more accessible.

Compared to traditional data, alternative data is generally more granular, observed more frequently, and remains useful for a shorter period of time. Since it comes mostly from external sources, it may seem riskier than data from your own business.

Let's look at some definitions and examples, and see how alternative data is being used today. Then we'll reconsider the risk of using – or not using – alternative data.

What’s the data like outside today?

Let’s start with a simple example. If you’re like me, you use your smartphone weather app a lot. This saves us from taking manual atmospheric readings throughout the day or rigging a bunch of meteorological sensors around the house and neighborhood. Instead, we simply open a weather app to quickly view alternative data about temperature, humidity, wind speed and precipitation. This granular data is only useful for a few hours at a time. For that reason, it needs to be updated frequently.

Now, imagine you run a retail business that sells umbrellas. Internal, traditional sales transaction data can't forecast rain, even though we know that people generally buy more umbrellas when it’s raining. But external, alternative data can forecast rain and other meteorological events using a continuous stream of climate data from satellites circling the planet and internet-connected sensors closer to the ground that constantly measure and transmit climate data.

The best way to process and analyze data of this volume and velocity is with automated algorithms that quickly crunch the numbers and update weather forecasts. This might not be as accurate as what a meteorologist could tell you, but it’s more likely to save your customers from getting caught in the rain without an umbrella.

That’s obviously an oversimplified example. Retail stores don't tend to make near-real-time stocking decisions based on weather forecasts. Property insurers, on the other hand, do use climate data when they assess how floods, hurricanes, tornadoes or wildfires could impact coverage. That's a practical business example of using this type of data.

Less risky alternatives

In the financial services industry, alternative data plays an increasing role in risk assessments. This is because traditional data, such as financial statements, quarterly earnings reports, SEC filings, management presentations and press releases are historical in nature. As such, they often don't provide enough information for investors to fully evaluate a company or an investment.

Banks and other financial institutions often need to look beyond their internal master and transaction data when attempting to better understand customers and prospects. Alternative data helps them supplement existing insights so they can:

  • Better evaluate their products and service offerings.
  • Effectively assess the economic landscape of regional and national marketplaces.

“The risk function is operating in a world with increasing demand for better digital services and more automated decision-making,” explained Terisa Roberts, Global Solution Lead for Risk Modeling and Decisioning at SAS. “Alternative data has proved its value in delivering insights beyond the scope of what traditional data provides. Especially where traditional data is considered limited, lagging or incomplete, alternative data can supplement that information.”

One example is the COVID-19 crisis. New types of data, such as online sales numbers and road traffic information, were put to good use during the global pandemic to approximate macroeconomic behavior. “Alternative data,” Roberts remarked, “gave us those insights and provided the necessary buttons for policymakers to push to make decisions on how to handle the pandemic.”

Other examples

You can use new, nontraditional data in many different ways. Here are just a few examples:

  • Individual behavior. Sentiment analysis relies on alternative data from social media and pattern analysis of web traffic and app usage.
  • Mobility tracking. Satellite images and geolocation data from internet-connected sensors and devices – as well as pings off public Wi-Fi networks and Bluetooth beacons – represent alternative data.
  • Economic evaluation. Government agencies, industry associations and third-party data providers offer periodic metrics that incorporate new data. Consider:
    • Jobless claims and new job listings.
    • House prices and mortgage rates.
    • Air travel bookings and hotel occupancy rates.
    • Restaurant reservations and movie ticket sales.
    • Online shopping price indices and eCommerce buying trends.
    • Retail revenue tracking via credit card transactions.

There are many benefits of applying alternative data to financial risk analysis. As Roberts explained, such data can supplement a risk analysis – for example, using phone, rent and utility payments to complement credit scoring.

More machine than manual

Since alternative data is more frequent, more granular and more time-sensitive, its use is more likely to identify temporary swings and pick up more nuanced behavior. It's also more likely to suffer from a higher noise-to-signal ratio. In addition, alternative data sources are often too large to manage and process manually. So, you may need to use machine learning and automation to generate meaningful and timely insights.

These new types of data also increase the speed at which you gain insights and make decisions. That's because alternative data sources are updated more frequently than traditional sources. Some financial institutions have seen tangible improvements from using alternative data – not only in terms of increased risk model accuracy but also in terms of delivering a better customer experience.

This implicit relationship with artificial intelligence (AI) is why many business users don't work directly with alternative data, or know much about the underlying processes that present its associated metrics. Alternative data is often unstructured big data of limited use in raw form. Organizations sometimes lack the IT infrastructure and expertise to handle this type of data. Instead, they frequently use the cloud to store alternative data and rely on AI and machine learning techniques to process and analyze it.

Getting a second opinion on the digital pulse

Alternative data is usually data that's external to the enterprise. This is why it can be so useful for taking the digital pulse of the overall economy or the specific sectors that affect your industry or business.

Just like an actual pulse, fluctuations in the data may or may not indicate that you need to act or make a decision. But by checking external data often – and combining it with traditional, internal data – you can get a more comprehensive, near-real-time picture of the constantly changing business world.

Just as with the monitors attached to a hospital patient, it's more effective to automate the measurement of digital business vital signs than to do it manually. You can even set up automated alarms to trigger human intervention.

Clearly, alternative data has many applications in the financial services industry, including improved decisioning. But all organizations should consider banking on alternative data to provide a second opinion when traditional data needs to be supplemented or complemented.

It might seem risky to rely on external data that's not completely controlled by your business. But in today’s digitally interconnected business world, it's actually riskier not to consider the alternative.

Learn practical steps banks can take to transform risk management

About Author

Jim Harris

Blogger-in-Chief at Obsessive-Compulsive Data Quality (OCDQ)

Jim Harris is a recognized data quality thought leader with 25 years of enterprise data management industry experience. Jim is an independent consultant, speaker, and freelance writer. Jim is the Blogger-in-Chief at Obsessive-Compulsive Data Quality, an independent blog offering a vendor-neutral perspective on data quality and its related disciplines, including data governance, master data management, and business intelligence.

Leave A Reply

Back to Top